TL;DR
A Microsoft Fabric workspace is the shared space where one team keeps its data, pipelines and reports. It also controls who can see and change that work. Teams split workspaces by owner, audience and release schedule. They keep development, test and production in separate workspaces. Compute and cost sit on the capacity, so capacity is a separate choice. Finished changes move to production one stage at a time, then reach business users through an app.
Key Takeaways Draw workspace boundaries by accountable owner, audience and release schedule, because those three decide who can change what. Keep development, test and production in separate workspaces, then link them with a deployment pipeline. Workspace design and capacity sizing are different decisions, since cost and compute sit on the capacity. Viewer is the only role that OneLake security default roles restrict, so data-level rules protect readers, not builders. Only an Admin can connect a workspace to Git, create its workspace identity or manage other Admins, so name two Admins early. On F64 or larger capacities, Free-license users with the Viewer role can read Power BI content. The Friday Release That Broke the Finance Report It is 4 p.m. on the last Friday of the quarter. A data engineer fixes a column mapping in a pipeline, then runs it to check the result. The pipeline sits in the same workspace as the finance team’s production semantic model, so the fix lands in production at once.
By Monday, the revenue report no longer matches the general ledger, so finance spends the morning tracing why. Nobody broke a rule, because the engineer held the Contributor role, and Contributors may edit and run pipelines. The design caused the failure, since one container held both development work and the numbers the CFO signs.
This guide first explains what a Microsoft Fabric workspace is and why it works as the boundary for ownership and release. It then shows how teams split, secure and run workspaces, so a Friday fix stays in development until someone promotes it.
What Is a Microsoft Fabric Workspace? A Microsoft Fabric workspace is a shared container for the items a team builds, such as lakehouses, warehouses, notebooks, pipelines, semantic models and reports. It also carries its own role assignments, a capacity assignment and a set of settings. As a result, everything inside it shares one access model and one source of compute.
Picture a finance analytics team. Its lakehouse, the pipelines that load it, the semantic model on top and the month-end reports all live in one workspace. Anyone who holds a role there can reach those items, while the capacity behind the workspace pays for the compute they use.
The workspace also sits in a clear hierarchy inside Microsoft Fabric . At the top, a tenant holds one or more capacities. Each capacity then hosts workspaces, and each workspace holds folders and items. Domains also sit beside that chain and group workspaces by business area, which helps with discovery and delegated administration.
Workspace vs Tenant, Capacity, Domain and Lakehouse These five terms get mixed up in planning meetings because each one draws a boundary. The real difference, then, is what each boundary controls and who owns the decision about it.
Concept What it is Who manages it Design decision it drives Tenant The organization’s whole Fabric environment, tied to Microsoft Entra ID Fabric administrator Tenant settings, such as who may create workspaces Capacity A pool of compute (F SKU, P SKU or trial) that workspaces draw on Capacity administrator Size, region and which workspaces share compute Domain A logical group of workspaces for one business area Fabric admin, then domain admins Discovery in the OneLake catalog and delegated settings Workspace The container for items, roles and settings Workspace Admin The boundary for ownership, access and release Lakehouse One item inside a workspace that stores files and Delta tables Item owner, within workspace roles Where data lands and how teams share it
A lakehouse is therefore one of the items a workspace contains. The guide to the Microsoft Fabric lakehouse covers how that item stores and serves data.
Is a Fabric Workspace Different From a Power BI Workspace? No, it is the same container. Fabric grew out of the Power BI service, so a workspace created in Power BI and one created in Fabric are the same object. What changes instead is the workspace type, which decides the items the New item menu offers.
Microsoft’s Fabric licensing page notes that license mode is now called workspace type, although the change is in name only. A Pro workspace holds Power BI items such as reports, dashboards and semantic models. A workspace on Fabric capacity can also hold lakehouses, warehouses, notebooks and pipelines.
Going back down is possible but not instant. The workspace creation guide says a move from a Premium type to Pro requires removing non-Power BI Fabric items first. The comparison of Microsoft Fabric vs Power BI also explains where each product fits.
Workspace Types and Capacity: What Decides What a Workspace Can Do A workspace’s type decides what you can build in it, and its capacity decides how much compute it gets. Both are settings on the workspace, but they answer different questions and often belong to different owners.
New workspaces start in shared capacity by default. An admin can then assign any workspace, including My workspace, to any capacity in the tenant. Cost follows the capacity, because F SKUs bill pay-as-you-go per second with a one-minute minimum. Splitting one workspace into five on the same capacity therefore adds no compute cost by itself.
Workspace type Family Runs on What it means for teams Pro Power BI Shared capacity, per-user Pro licenses Power BI items only (reports, dashboards, semantic models) Premium Per User Power BI Per-user PPU licenses Power BI items for teams licensed per user Embedded Power BI A SKUs Power BI content embedded in your own applications Fabric Fabric F SKUs, billed per second Every Fabric item, plus free Viewers on F64 or larger Fabric Trial Fabric Trial capacity of 64 CU for 60 days Every Fabric item, for evaluation before you buy Power BI Premium Fabric P SKUs Fabric items on an existing Premium capacity
The free-viewer rule matters most when a team publishes widely. On F64 or larger, users with a Free license and the Viewer role can read Power BI content. Below F64, every consumer needs a Pro or PPU license, and creating Power BI items outside My workspace still needs Pro.
Sizing the capacity itself is another exercise, covered in Fabric capacity planning . For the meters behind the bill, the Fabric pricing guide walks through each one.
Kanerika Service
Microsoft Fabric Consulting and Implementation
Kanerika designs Fabric estates end to end, from workspace and capacity design to governance, pipelines and Power BI, for teams moving analytics onto one platform.
Explore Microsoft Fabric Services Workspace Roles: Admin, Member, Contributor and Viewer Roles decide what each person can do inside a workspace, and Fabric has exactly four of them. A role applies to every item in the workspace at once, which is why the workspace boundary matters so much for security.
The workspace roles reference lets you assign a role to users, security groups, Microsoft 365 groups, distribution lists or service principals. Groups are the better choice, especially at scale. A new analyst then gains access by joining a group, instead of filing a ticket with every workspace Admin.
Each role adds powers on top of the one below it, so the ladder runs from reading content up to managing people. The table lists the capabilities that cause the most confusion in practice.
Capability Roles that can do it Update or delete the workspace Admin only Add or remove people, including other Admins Admin only Connect the workspace to Git Admin only Create the workspace identity Admin only Add Members or lower roles Admin and Member Allow others to reshare items Admin and Member Create and modify warehouse and database items Admin, Member and Contributor Write, delete and run pipelines, notebooks and Spark jobs Admin, Member and Contributor Read all data through OneLake APIs or Spark Admin, Member and Contributor View content and query the SQL analytics endpoint All four roles, including Viewer
Three rows matter more than the rest, especially in a new estate. Only an Admin can connect Git or create the workspace identity, so those steps always need the platform owner. A Contributor can write and run pipelines and notebooks but cannot add people, which is exactly the access level in the Friday release story.
What the Viewer Role Can and Cannot See The Viewer row is subtle, because it behaves differently from the others. A Viewer reads data through the SQL analytics endpoint rather than through Spark. Under the OneLake data access control model , it is also the only role that OneLake security default roles restrict. A ReadWrite setting on Admin, Member or Contributor has no effect, since those roles already have full access. Row- and column-level rules therefore protect readers, while builders see everything in the workspace.
A Role Worksheet by Persona A worksheet then turns the four roles into a staffing decision. Start from the job each person does, then give the narrowest role that still lets them do it.
Persona Role Why Watch-out Platform owner Admin Connects Git, creates the workspace identity, manages people Name at least two Admins so access never depends on one person Data engineer Contributor in dev, Viewer in production Builds and runs pipelines and notebooks Production changes arrive through the pipeline, not by hand BI developer Contributor in dev and test Builds semantic models and reports Promotion to production goes through the release owner Analyst Viewer Reads reports and queries the SQL endpoint Needs Contributor only in a sandbox workspace Business consumer No workspace role, app audience only Sees published content Keeps consumers out of the working area Service principal Contributor in each workspace it automates or deploys to Runs automation and deployments Avoid Admin unless it must manage people
Assign every role to a Microsoft Entra group, then manage membership in Entra. The guide to Azure identity management covers group design, and data access governance explains how to review who holds which access over time.
How Teams Use Fabric Workspaces Day to Day Teams use workspaces to separate who builds, who owns and who consumes. A healthy estate reads like a supply chain, where data moves from engineering to modeling to consumption, with an owner at every stage.
Two splits run at once. By team, engineering and BI each own their work, and by environment, each team’s work moves from development to production.
Ingest in an engineering workspace. Data engineers land source data in a lakehouse with pipelines and notebooks, then clean it.Model in a BI workspace. BI developers build semantic models and reports on curated tables, often with Direct Lake semantic models .Develop in dev. Each team’s changes start in its own development workspace, connected to Git.Promote to test and production. A deployment pipeline then moves approved items one stage at a time.Publish to consumers through an app. Business users read from an app, never from the working area.Here is how that looks when two teams share data. Consider a manufacturer with a finance team and an operations team that need the same conformed customer and product tables. They report to different leaders and release on different days, so one shared production workspace would tie them together.
In practice, the platform team keeps the conformed tables in one engineering workspace. Each business team then reads them through OneLake shortcuts instead of copying them, so there is still one copy of the truth.
Finance gets its own dev, test and production workspaces, and so does operations. A finance release on quarter-end Friday cannot touch an operations report, because the two never share a production workspace. Each team also publishes its own app, so each audience sees only its own content.
How Many Workspaces? Four Deployment Patterns There is no fixed number of workspaces, but Microsoft’s Architecture Center describes four Fabric deployment patterns that frame the choice. Each one trades simplicity for isolation, so the right pick depends on how many owners you have.
Pattern When it fits Trade-off Monolithic: one workspace for all use cases and teams A small organization with one central data team Simple to run, but dev and production collide and access is all or nothing Multiple workspaces on one capacity Several teams or environments sharing one budget Clean ownership, though a busy workload can slow others on that capacity Multiple workspaces on separate capacities Teams with different budgets, regions or performance needs Isolated compute and clear chargeback, with more capacities to manage Multiple tenants Legal or regulatory separation between entities The strongest isolation, with separate administration in each tenant
For a mid-sized estate, the second pattern is the usual starting point, and the third arrives once chargeback or regional rules appear. Region choices and disaster recovery belong to Fabric architecture and region design , which covers them in depth.
Five Tests Before You Add Another Workspace A new workspace should only exist when it solves a real boundary problem. Run these five tests first, and use a folder when none of them is true.
Is a different person accountable for the content? Does it serve a different audience or need different access? Will it release on a different schedule? Does it need a different capacity or region? Would keeping it here push the workspace toward the 1,000-item limit? A single yes justifies a new workspace. If every answer is no, a folder inside the existing workspace keeps items organized without adding another role list to maintain.
Talk to Kanerika
Planning Your Workspace Boundaries?
Kanerika can review your owners, audiences and release schedules and map them to a workspace and capacity plan. Book a short working session with the team.
Book a Meeting → Do Medallion Layers Need Separate Workspaces? Only when the people change. Split bronze and silver from gold when different owners maintain them or different audiences read them. For example, a platform team might own raw ingestion while a business team owns the curated gold tables.
Otherwise, keep the layers together and separate by environment instead. One lakehouse per layer inside one workspace is easier to run than three workspaces with identical role lists.
How to Create a Microsoft Fabric Workspace (and What to Set First) Creating a workspace takes a few clicks, but a few choices on the creation screen are hard to change later. Decide them before you select Apply, using the steps from Microsoft’s guide to creating workspaces .
Select Workspaces in the navigation pane, then New workspace . Enter a name. It must be unique, so Fabric rejects a name already in use. Add a description so others know what the workspace is for. Assign a domain if your organization uses domains, although this step is optional. Open Advanced to set the contact list, the workspace type and the default storage format for semantic models. Leave the template app option off unless you are building a template app, because that choice cannot be reverted. Select Apply . Once it opens, the workspace shows its items in a list view. A lineage view also traces how data flows from sources to reports. Folders group related items without changing their access, and the New item menu lists only what the workspace type allows.
Watch on YouTube
Microsoft Fabric Setup Guide: 5 Budget-Killing Mistakes to Avoid
A short clip on the setup choices that inflate Fabric spend, worth a watch before you assign a new workspace to a capacity.
Settings to Review After Creation Before anyone builds, an Admin should review the settings below. That is because changing workspace settings counts as updating the workspace, which the roles reference reserves for Admins.
Setting What it controls Git integration Version control and branching for workspace items Workspace identity A managed service principal for trusted access to secured sources Spark settings Default runtime, pools and environment for notebooks and jobs OneLake settings Workspace-level options for how its OneLake data is reached Microsoft 365 and SharePoint A link to a Microsoft 365 group’s SharePoint site Azure connections Links to Azure resources, such as an ADLS Gen2 account Monitoring Workspace monitoring that writes logs to an Eventhouse
Connect Git on day one, because version history starts only from the first commit. Fill in the contact list too, because a failed refresh needs someone to notify. Everything else can wait until after the first build. For the wider rollout plan, Kanerika’s Microsoft Fabric Checklist puts these early decisions in context.
Checklist
Microsoft Fabric Checklist
Practical guidance for planning a Fabric rollout across architecture, migration, governance and security, so workspace decisions like these fit the wider plan.
Get the Checklist → Releasing Changes: Git Integration and Deployment Pipelines Releasing safely means changes reach production in a known order, with a record of what changed. Fabric gives each workspace two tools for that, Git integration for version control as well as deployment pipelines for promotion.
Git integration works at the workspace level, so each workspace has its own connection. An Admin connects the workspace to a branch in Azure DevOps, GitHub or GitHub Enterprise, cloud versions only. Fabric then keeps the folder structure in the repository. Support varies by item type, and some items are still in preview, so check the list before you rely on it.
Branched workspaces also add isolation for each developer. When a developer needs it, they branch out from the shared dev workspace into a workspace linked to a feature branch. They build there, then merge back through a pull request.
Watch on YouTube
Microsoft Fabric Deployment Pipelines: ALM Setup & 2026 Deadlines
A walkthrough of how deployment pipelines move Fabric content from development to production, the release discipline that keeps a Friday fix out of live reports.
Deployment pipelines then promote content. A pipeline has between 2 and 10 stages, with 3 by default, and each stage is a workspace. Fabric pairs items between stages, so a deployment updates the matching item in the next stage instead of creating a copy.
In the Friday release story, this setup keeps the column fix in dev until a reviewer approves it and someone deploys it to test. The guide to Fabric deployment pipelines covers setup in more depth than this summary. For report-only teams, the Power BI deployment pipelines guide shows the same mechanics.
Securing a Workspace: Identity, OneLake Security and Domains Workspace roles are the first security layer, but three more sit around them. Each one answers a different question. Who is the workspace when it calls other services, what may a reader see, and which business area owns it?
Workspace identity. An automatically managed service principal tied to the workspace, per Microsoft’s workspace identity documentation . It is not available in My workspace, but it works on any capacity SKU. Trusted access to firewall-enabled ADLS Gen2 accounts, however, needs a purchased F SKU.OneLake security roles. Row- and column-level rules on lakehouse data that bind Viewers. Check the feature’s current release status before you plan on it, and see OneLake security roles for design patterns.Domains. Fabric admins create domains and then assign workspaces to them. Domain admins and contributors then manage settings for their area, and subdomains add a further level.A practical order is roles first, then identity, then data-level rules. Domains come last, because they organize workspaces that already have clean boundaries.
Tenant-wide policy sits above all of this. The guide to Fabric governance from day one covers those settings, while Microsoft Purview extends cataloging and labeling across the wider estate. For discovery beyond Fabric, an enterprise data catalog plays a similar role, especially across mixed platforms.
Monitoring, Limits and Recovery Running a workspace well means knowing what ran, what it cost and how to undo a mistake. The table sums up the behaviors a platform owner should know before go-live.
Two hard limits are also worth learning early. According to Microsoft’s workspace documentation , one workspace can hold up to 1,000 Fabric and Power BI items. A single user or service principal can also be a member of up to 1,000 workspaces.
Item How it behaves Workspace monitoring Preview. Sends logs and metrics to an Eventhouse in a monitoring item, keeps 30 days by default, uses capacity, and its database counts toward the item limit Recycle bin Preview. Item Recovery is on by default, the tenant admin sets retention from 3 to 90 days, and only workspace Admins delete permanently Item limit Up to 1,000 Fabric and Power BI items per workspace Membership limit A user or service principal can belong to up to 1,000 workspaces
Workspace monitoring is opt-in, and once on, it draws on the same capacity as your workloads. Turn it on where you need query and job history, rather than everywhere by habit. The roundup of data pipeline monitoring tools compares options outside Fabric.
The recycle bin also changes how teams handle a bad delete. An item removed by mistake can come back within the retention window. The first call after an accident therefore goes to a workspace Admin, not to a backup team.
Spark and Airflow Settings at the Workspace Level Data engineering workspaces carry their own Spark and orchestration defaults. Two of them deserve a check before anyone builds in a new workspace.
The Spark runtime comes first. Microsoft’s Fabric runtime page lists Runtime 2.0 as generally available, with Spark 4.1, Delta Lake 4.2, Java 21 and Python 3.13, and recommends it for production. Runtime 1.3, built on Spark 3.5.5 and Delta Lake 3.2, has an end-of-support announcement. New workspaces should default to 2.0, while older ones need a migration plan.
Orchestration is the second. An Apache Airflow job is the Fabric successor to the Workflow Orchestration Manager in Azure Data Factory. It does not support private or virtual networks yet, so plan orchestration differently for workspaces that depend on network isolation.
The deeper guide to Fabric data engineering covers runtimes and notebooks in more detail than this section. For a wider comparison of schedulers, including open-source Apache Airflow , see these workflow orchestration tools .
Common Workspace Design Mistakes Workspace problems usually trace back to a few design choices made in the first week of a Fabric rollout. Each one is cheap to avoid early and expensive to unwind later.
Most of these mistakes stay hidden during a first pilot. They only bite once the estate grows, and more teams start to share the same workspaces and capacity. Microsoft MVP Amit Chandak talks through why Fabric deployments break down at scale in this episode of the Kanerika podcast.
Mistake Symptom Fix Dev and production in one workspace Untested edits reach production reports Separate dev, test and production workspaces linked by a pipeline Roles given to individuals Access reviews take days and leavers keep access Assign every role to an Entra group Everyone made Contributor Anyone can change or run production items Viewer by default, Contributor only in dev One workspace per report Hundreds of workspaces with copied role lists Group by owner and audience, then use folders Splitting a workspace to fix slow queries No speed gain, more administration Check the capacity, since compute sits there No owner or contact list Nobody to call when a refresh fails Fill in the contact list at creation Git connected after go-live No history for the first releases Connect Git before the first build
The performance row deserves a second look, because it is the least intuitive. A workspace does not own compute, so moving items into a new workspace on the same capacity changes nothing about speed. The fix lives in capacity sizing or in the workload itself, for example a heavy refresh scheduled at peak hours.
Approve Before Production Before a workspace goes live, a short approval list catches the rest. Treat each item as a gate that the platform owner signs off before release.
Separate dev, test and production workspaces exist for the workload. Every role is assigned to an Entra group, with at least two Admins. The workspace sits on the intended capacity and region. Git is connected, and the main branch holds the current version. A deployment pipeline links the stages, and item pairing is confirmed. A workspace identity is in place for any firewall-protected source. Monitoring, recycle-bin retention and the contact list are set. If any item is still open, the workspace stays in test until someone closes it.
How Kanerika Designs Fabric Workspace Estates Kanerika treats a workspace estate as an operating model rather than a folder structure. As a Microsoft partner , the team designs workspaces as part of its Microsoft Fabric services , in four short steps.
The steps run in order, because each one produces something the next one needs.
Assess the existing estate. Inventory workspaces and items, name owners, measure role sprawl and map workspaces to capacities.Define boundaries and access. Split workspaces by owner, audience and release, then assign roles through Entra groups.Prove it on one real workload. Connect Git and run a three-stage deployment pipeline before scaling out.Hand over with operations in place. Set up workspace monitoring, a recycle-bin policy and a quarterly access review.Applied to the Friday release story, step two is where dev and production would have been split, and step three is where the pipeline would have stopped the column fix before it reached the CFO’s numbers.
Forest River: A Review Before Go-Live The same lens also applies to estates that already exist. For Forest River, Kanerika ran a six-week, read-only architecture review of the Fabric workspace, 124-table Dataverse mirror, pipelines, notebooks and security settings. It took place before a Dynamics 365 go-live.
The review produced about 136 findings, scored on a RAG scorecard across five lenses and 18 dimensions. Pre-cutover actions were scoped at an estimated 30 days, and the result was a conditional go-live recommendation with target states for DevOps, security and governance. The Forest River architecture review describes the method.
Teams earlier in the journey can start with a Fabric adoption roadmap or book a working session to review their own workspace plan.
Case Study
Fabric Architecture Review Before a Dynamics 365 Go-Live
A six-week, read-only review of a Fabric workspace, a 124-table Dataverse mirror, pipelines, notebooks and security settings produced about 136 findings and a conditional go-live recommendation.
Read the Case Study → So a Fabric workspace is where a team owns, secures and releases its work, and it works best when split by owner, audience and release.
Frequently Asked Questions
What is a workspace in Microsoft Fabric? A workspace in Microsoft Fabric is a container for the items a team builds, such as lakehouses, warehouses, notebooks, pipelines, semantic models and reports. It carries its own roles, capacity assignment and settings, so everything inside shares one access model. Teams use workspaces as the boundary for ownership, access and release, rather than as simple folders.
What is a workspace role in Microsoft Fabric? A workspace role sets what a person or group can do across every item in a workspace. Fabric has four roles, Admin, Member, Contributor and Viewer, and each adds powers to the one below. You can assign them to users, security groups, Microsoft 365 groups, distribution lists or service principals, and groups are easier to manage.
What is the difference between a workspace and a lakehouse? A lakehouse is one item inside a workspace. It stores files and Delta tables and exposes a SQL analytics endpoint for queries. The workspace around it holds the roles, capacity assignment and settings that govern the lakehouse and every other item. One workspace can hold several lakehouses, for example one for each medallion layer of a data estate.
Which license do I need to create a Fabric workspace? Your tenant admin decides who may create workspaces, and the items inside them carry their own license needs. Creating Power BI items outside My workspace requires a Pro license. Fabric items such as lakehouses and notebooks require the workspace to sit on a Fabric capacity, a Power BI Premium capacity or an active Fabric trial.
How many workspaces should a team have? Start with as few as the boundaries require. A typical team runs separate dev, test and production workspaces per workload, and adds more only when the owner, audience, release schedule, capacity or region differs. Use folders for everything else. Microsoft’s four deployment patterns, from monolithic to multi-tenant, help frame the choice for larger estates.
Should I separate dev, test and production workspaces? Yes, for any workload that serves real users. Separate workspaces stop an untested edit from reaching production reports, because Contributors can edit and run items wherever they hold the role. Link the three workspaces with a deployment pipeline, which supports 2 to 10 stages and pairs items so each deployment updates the matching item downstream.
Do bronze, silver and gold layers need separate workspaces? Only when different people own or read them. Split bronze and silver from gold when a platform team owns ingestion and a business team owns curated tables, or when audiences differ. Otherwise keep the layers in one workspace as separate lakehouses and separate by environment instead, which avoids maintaining three identical role lists.
Can multiple workspaces share one Fabric capacity? Yes. One capacity can host many workspaces, which is one of Microsoft’s standard deployment patterns. The workspaces share compute, so a heavy workload in one can slow the others. Adding another workspace does not improve performance, because the capacity governs compute. Separate capacities are the answer when isolation or chargeback matters to the business.
Can I move a workspace to another capacity? Yes. An admin can assign any workspace, including My workspace, to any capacity in the tenant, and can reassign it later. Moving down from a Premium or Fabric type to Pro requires removing non-Power BI Fabric items first. Check capacity size and region before a move, since the new capacity supplies all compute afterwards.
How do Git integration and deployment pipelines work with workspaces? Git integration connects one workspace to a branch in Azure DevOps, GitHub or GitHub Enterprise, and only an Admin can set it up. Developers can branch out into their own workspace for isolated work. Deployment pipelines then promote content through 2 to 10 stages, each one a workspace, with items paired so updates land correctly.
What is a workspace identity in Microsoft Fabric? A workspace identity is an automatically managed service principal tied to one workspace. Only an Admin can create it, and it is not available in My workspace. It works on any capacity SKU, though trusted access to firewall-enabled ADLS Gen2 storage needs a purchased F SKU. Teams use it to reach secured sources without personal credentials.
How do I monitor activity in a Fabric workspace? Turn on workspace monitoring, a preview feature that sends logs and metrics to an Eventhouse inside a monitoring item. Data stays for 30 days by default. Monitoring consumes capacity, and its database counts toward the workspace item limit, so enable it where you need query and job history rather than in every workspace by habit.
How do you create a Microsoft Fabric workspace? Select Workspaces in the navigation pane, then New workspace. Enter a unique name, add a description and optionally assign a domain. Under Advanced, set the contact list, workspace type and default storage format for semantic models. Leave the template app option off unless you need it, because it cannot be reverted, then select Apply.
What happens when you delete an item in a Fabric workspace? The item goes to the recycle bin, a preview feature that is on by default through the Item Recovery tenant setting. The tenant admin sets retention between 3 and 90 days, and a deleted item can be restored at any point during that window. Only workspace Admins can delete an item permanently before the retention period ends.
How many items can a Fabric workspace hold? A workspace can hold up to 1,000 Fabric and Power BI items, and a single user or service principal can belong to up to 1,000 workspaces. Monitoring databases count toward the item limit too. If a workspace approaches the limit, split it along an owner, audience or release boundary rather than by item type.
Can users view a report without workspace access? Yes. Most business users should read content through a workspace app or through items shared with them, not through a workspace role. Licensing still applies to every consumer. Microsoft’s licensing page says Free-license users with the Viewer role can view Power BI content on F64 or larger, while smaller capacities need Pro or PPU licenses.
What is the best practice for Microsoft Fabric workspace design? Draw boundaries by accountable owner, audience and release schedule. Keep development, test and production in separate workspaces linked by a deployment pipeline, and assign roles to Entra groups instead of individuals. Connect Git before the first build, fill in the contact list at creation, and use folders rather than new workspaces when no boundary differs.
What is the difference between a Fabric domain and a workspace? A workspace holds items, roles and settings for one team or workload. A domain is a logical group of workspaces for a business area, such as finance or sales. Fabric admins create domains and assign workspaces to them, and domain admins then manage settings for their area. Domains also organize content in the OneLake catalog.
Is a Microsoft Fabric workspace free to use? Creating a workspace has no charge, and new workspaces start in shared capacity by default. A shared-capacity workspace supports Power BI items, and creating them outside My workspace needs a Pro license. Fabric items such as lakehouses need a Fabric capacity, while a Fabric trial gives 64 capacity units for 60 days before you buy.
How much does a Fabric workspace cost to run? A workspace has no price of its own. Cost sits on the capacity the workspace is assigned to, and F SKUs bill pay-as-you-go per second with a one-minute minimum at regional prices. Several workspaces on one capacity share that compute, so splitting a workspace adds no compute cost by itself. Creator and consumer licenses still apply.
Is Microsoft Fabric part of Microsoft 365? No. Microsoft Fabric is licensed separately from Microsoft 365, through capacities and Power BI licenses. The two still connect well. A workspace can link to a Microsoft 365 group’s SharePoint site, and Microsoft 365 groups can hold workspace roles. That lets teams reuse the groups and collaboration sites they already run when they grant Fabric access.
What is the difference between a Fabric workspace and a Power BI workspace? There is no separate object. Fabric grew out of the Power BI service, so both are the same workspace. The workspace type, formerly called license mode, decides which items you can create. A Pro workspace holds Power BI items such as reports, dashboards and semantic models. A workspace on Fabric capacity can also hold lakehouses, warehouses, notebooks and pipelines.
What is the difference between a workspace and a capacity in Fabric? A workspace is a container for items, roles and settings. A capacity is a pool of compute that workspaces draw on, sized and billed as an F SKU, a P SKU or a trial. One capacity can host many workspaces, and an admin can reassign a workspace to another capacity. Design shapes workspaces, while budget shapes capacity.
